pith. sign in

arxiv: 2603.25099 · v2 · pith:XGUY3L4Knew · submitted 2026-03-26 · 💻 cs.CE · cs.AI

Large Language Models as Optimization Controllers: Adaptive Continuation for SIMP Topology Optimization

Shaoliang Yang , Jun Wang , Yunsheng Wang This is my paper

Pith reviewed 2026-05-19 17:29 UTC · model grok-4.3

classification 💻 cs.CE cs.AI
keywords large language modelstopology optimizationSIMP methodadaptive continuationstructural optimizationparameter controlcompliance minimization
0
0 comments X

The pith

A large language model can act as a real-time controller for SIMP topology optimization by choosing parameters from current state observations instead of a fixed schedule.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that an LLM can replace conventional fixed-schedule continuation in SIMP topology optimization with state-conditioned decisions at regular intervals. The model receives a short vector of compliance, grayness, stagnation, checkerboard measure, volume fraction, and budget use, then outputs new values for the penalization exponent, projection sharpness, filter radius, and move limit. On three 2-D and two 3-D benchmark problems the LLM-controlled runs reach the lowest compliance of all tested methods, improving 5.7 to 18.1 percent over a no-continuation baseline while producing fully binary designs. An ablation that keeps only the schedule without the LLM underperforms the fixed baseline on two problems, indicating that the live interventions drive the gain. A meta-optimization loop and a grayness gate keep the process stable across runs.

Core claim

An LLM supplied with a structured observation vector at every k-th iteration can output numerical settings for p, β, r_min, and δ that produce lower final compliance than fixed continuation, three-field continuation, expert heuristics, or schedule-only ablation on all five benchmark geometries, while guaranteeing fully binary final designs after a standardized sharpening tail.

What carries the argument

Direct Numeric Control interface in which the LLM maps the six-element observation vector to updated optimization parameters, guarded by a hard grayness gate and tuned by a second LLM meta-optimization loop.

If this is right

  • Topology-optimization workflows can become fully automatic without hand-crafted continuation schedules.
  • All final designs remain binary without extra post-processing steps.
  • The same observation-plus-control pattern could be tested on other iterative engineering solvers that currently rely on fixed parameter ramps.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The approach opens a path to letting language models steer other numerical optimization loops whose state can be summarized in a compact numeric vector.
  • If the grayness gate and meta-optimization steps are removed, performance may drop on problems that require aggressive early binarization.

Load-bearing premise

The LLM's internal mapping from the supplied observation vector to parameter values is reliable enough to beat both fixed rules and expert heuristics in real time.

What would settle it

Run the identical LLM controller on a new 2-D or 3-D problem with the same resolution and iteration budget; if its final compliance is not lower than the expert-heuristic baseline, the advantage claim does not hold.

Figures

Figures reproduced from arXiv: 2603.25099 by Jun Wang, Shaoliang Yang, Yunsheng Wang.

Figure 1
Figure 1. Figure 1: Overall two-level system pipeline. The inner loop executes the SIMP solver for up to [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Three-field SIMP formulation. Raw design variables [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: LLMagent decision loop at iteration 147 (49% of budget consumed). [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Meta-optimization outer loop. After each completed comparison run, the meta-optimizer [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Final compliance distributions (n = 5 runs) for all five controllers. Each panel plots the mean (filled circle) and ±3σ interval (horizontal bars). The LLM agent achieves the lowest mean on every problem with narrow spread, confirming that gains are reproducible across random seeds. The schedule-only controller underperforms the fixed no-continuation baseline on cantilever (+0.43%) and MBB beam (+1.09%), d… view at source ↗
Figure 6
Figure 6. Figure 6: Final physical density fields ˜ρ and element-density histograms for all five controllers (120×60, Vf = 0.40, representative run). The fixed controller retains substantial gray material (G > 0), visible as blurred member boundaries and a non-bimodal histogram. All continuation controllers achieve G = 0.000 with bimodal histograms concentrated at ˜ρ ∈ {0, 1}. 18 [PITH_FULL_IMAGE:figures/full_fig_p018_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Compliance convergence curves for the cantilever benchmark (120 [PITH_FULL_IMAGE:figures/full_fig_p019_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Grayness convergence G(t) for the cantilever benchmark (120×60, representative run). The dashed horizontal line marks the grayness gate threshold (G = 0.20). The fixed controller plateaus at G ≈ 0.10 and never achieves full binarization. All continuation controllers eventually reach G = 0.000; the LLM agent crosses the gate threshold around iteration 150—substantially later than three-field continuation (∼… view at source ↗
Figure 9
Figure 9. Figure 9: Hyperparameter trajectories for all five controllers on the cantilever benchmark (120 [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: 3-D density projections (X-ray view) for the 40 [PITH_FULL_IMAGE:figures/full_fig_p024_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: 3-D density projections (X-ray view) for the 40 [PITH_FULL_IMAGE:figures/full_fig_p025_11.png] view at source ↗
read the original abstract

We present a framework in which a large language model (LLM) acts as an online adaptive controller for SIMP topology optimization, replacing conventional fixed-schedule continuation with real-time, state-conditioned parameter decisions. At every $k$-th iteration, the LLM receives a structured observation$-$current compliance, grayness index, stagnation counter, checkerboard measure, volume fraction, and budget consumption$-$and outputs numerical values for the penalization exponent $p$, projection sharpness $\beta$, filter radius $r_{\min}$, and move limit $\delta$ via a Direct Numeric Control interface. A hard grayness gate prevents premature binarization, and a meta-optimization loop uses a second LLM pass to tune the agent's call frequency and gate threshold across runs. We benchmark the agent against four baselines$-$fixed (no-continuation), standard three-field continuation, an expert heuristic, and a schedule-only ablation$-$on three 2-D problems (cantilever, MBB beam, L-bracket) at $120\!\times\!60$ resolution and two 3-D problems (cantilever, MBB beam) at $40\!\times\!20\!\times\!10$ resolution, all run for 300 iterations. A standardized 40-iteration sharpening tail is applied from the best valid snapshot so that compliance differences reflect only the exploration phase. The LLM agent achieves the lowest final compliance on every benchmark: $-5.7\%$ to $-18.1\%$ relative to the fixed baseline, with all solutions fully binary. The schedule-only ablation underperforms the fixed baseline on two of three problems, confirming that the LLM's real-time intervention$-$not the schedule geometry$-$drives the gain. Code and reproduction scripts will be released upon publication.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes an LLM-based adaptive controller for SIMP topology optimization that replaces fixed continuation schedules with real-time, state-conditioned updates to the penalization exponent p, projection sharpness β, filter radius r_min, and move limit δ. At every k-th iteration the LLM receives a six-element observation vector (compliance, grayness index, stagnation counter, checkerboard measure, volume fraction, budget consumption) and outputs numerical parameter values through a Direct Numeric Control interface; a hard grayness gate prevents premature binarization and a second LLM pass performs meta-optimization of k and the gate threshold. The agent is benchmarked against fixed, three-field continuation, expert-heuristic, and schedule-only ablation baselines on three 2-D and two 3-D problems (300 iterations, standardized 40-iteration sharpening tail), reporting 5.7–18.1 % lower final compliance and fully binary designs on every instance, with the ablation study cited as evidence that real-time intervention rather than schedule geometry drives the improvement.

Significance. If the performance advantage can be reproduced under fair hyperparameter protocols and with statistical controls, the work would demonstrate a practical route for embedding LLMs as online controllers in established topology-optimization pipelines, potentially reducing manual schedule design while preserving the interpretability of the underlying SIMP formulation. The multi-problem benchmark (2-D and 3-D) and the explicit schedule-only ablation constitute positive methodological steps that strengthen the empirical case.

major comments (3)
  1. [Experimental protocol and ablation study (abstract and §4)] The meta-optimization loop that tunes call frequency k and grayness-gate threshold is applied only to the LLM agent; the fixed, three-field, expert-heuristic, and schedule-only ablation baselines receive no equivalent search. Because the schedule-only ablation already underperforms the fixed baseline on two of the three 2-D problems, any reported gap may partly reflect the agent’s extra hyperparameter budget rather than the real-time mapping from the six-element observation vector. This directly undermines the central attribution that “the LLM’s real-time intervention—not the schedule geometry—drives the gain.”
  2. [Results section and Table 1 (or equivalent benchmark table)] No error bars, standard deviations, or statistical tests accompany the reported compliance reductions (−5.7 % to −18.1 %). With only single-run results per configuration and no repeated trials, it is impossible to assess whether the observed differences exceed run-to-run variability inherent to the stochastic elements of the LLM calls and the optimization itself.
  3. [Ablation study and discussion of attribution] The claim that the structured observation vector plus the LLM’s learned mapping is sufficient for superior decisions rests on the schedule-only ablation underperforming the fixed baseline; however, the ablation itself is not given the same meta-optimization treatment, leaving open the possibility that a well-tuned fixed schedule could close the gap without any LLM involvement.
minor comments (2)
  1. [Methods / Implementation details] Prompt templates, exact LLM model version, temperature settings, and the precise format of the Direct Numeric Control output are not provided; these details are necessary for reproducibility even if code is released later.
  2. [Abstract and reproducibility statement] The manuscript states that “Code and reproduction scripts will be released upon publication,” but does not include a current repository link or a minimal working example; this should be supplied at submission for a computational paper.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. The comments highlight important aspects of experimental fairness and statistical robustness that we will address in the revision. We respond to each major comment below.

read point-by-point responses

  1. Referee: The meta-optimization loop that tunes call frequency k and grayness-gate threshold is applied only to the LLM agent; the fixed, three-field, expert-heuristic, and schedule-only ablation baselines receive no equivalent search. Because the schedule-only ablation already underperforms the fixed baseline on two of the three 2-D problems, any reported gap may partly reflect the agent’s extra hyperparameter budget rather than the real-time mapping from the six-element observation vector. This directly undermines the central attribution that “the LLM’s real-time intervention—not the schedule geometry—drives the gain.”

    Authors: We agree that the meta-optimization was applied exclusively to the LLM agent, which introduces an asymmetry in hyperparameter effort. To correct this, we will extend a comparable search over fixed continuation schedules for the schedule-only ablation (optimizing parameters such as p and β progression rates). The revised results will be reported in §4 and the abstract, allowing a direct test of whether the performance gap persists under equivalent tuning budgets. This will strengthen the attribution to real-time state-conditioned decisions. revision: yes

  2. Referee: No error bars, standard deviations, or statistical tests accompany the reported compliance reductions (−5.7 % to −18.1 %). With only single-run results per configuration and no repeated trials, it is impossible to assess whether the observed differences exceed run-to-run variability inherent to the stochastic elements of the LLM calls and the optimization itself.

    Authors: We acknowledge that single-run results limit the ability to quantify variability. In the revised manuscript we will perform five independent replications of each configuration (including all baselines), reporting mean compliance, standard deviation, and a brief note on sources of stochasticity from LLM sampling. This will be added to the results section and Table 1. revision: yes

  3. Referee: The claim that the structured observation vector plus the LLM’s learned mapping is sufficient for superior decisions rests on the schedule-only ablation underperforming the fixed baseline; however, the ablation itself is not given the same meta-optimization treatment, leaving open the possibility that a well-tuned fixed schedule could close the gap without any LLM involvement.

    Authors: This concern is closely related to the first comment. By applying meta-optimization to the schedule-only ablation as described above, we will directly evaluate whether an optimized fixed schedule can match or exceed the LLM agent. Updated discussion text will explicitly address this possibility and interpret the new results in terms of the value of real-time adaptation versus schedule geometry alone. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical performance claims rest on external benchmark comparisons

full rationale

The paper describes an LLM controller framework for SIMP topology optimization and evaluates it through direct empirical runs against four independent baselines (fixed, three-field continuation, expert heuristic, schedule-only ablation) on standardized 2-D and 3-D problems. No mathematical derivation chain, first-principles prediction, or fitted parameter is presented that reduces by construction to its own inputs; compliance differences are measured after a fixed 40-iteration sharpening tail, and the schedule-only ablation is used explicitly to isolate real-time intervention effects. The meta-optimization loop is part of the proposed agent but does not alter the external measurement of outcomes against untuned baselines. The work is self-contained against these benchmarks with no self-citation load-bearing or self-definitional steps.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The framework depends on empirical tuning of call frequency and grayness gate threshold via a second LLM meta-optimization pass, plus the domain assumption that the chosen state metrics suffice for effective control.

free parameters (2)
  • LLM call frequency k
    Interval between LLM interventions, tuned by meta-optimization loop.
  • grayness gate threshold
    Value preventing premature binarization, tuned by second LLM pass.
axioms (2)
  • domain assumption SIMP topology optimization converges to binary designs under appropriate continuation of penalization and projection parameters
    Invoked throughout the benchmark setup and sharpening tail procedure.
  • ad hoc to paper The LLM can reliably map the six-element observation vector to parameter values that improve exploration over fixed or heuristic schedules
    Core premise of the adaptive controller and the source of claimed gains.

pith-pipeline@v0.9.0 · 5859 in / 1475 out tokens · 63154 ms · 2026-05-19T17:29:57.153375+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. IterSIMP-{\sigma}: Evaluating LLM-Assisted Spatial Interventions in Stress-Aware Topology Optimization

    cs.CE 2026-05 unverdicted novelty 5.0

    IterSIMP-σ integrates multimodal LLMs for proposing spatial density interventions in stress-aware SIMP topology optimization, yielding comparable but statistically non-significant performance gains over rule-based bas...

Reference graph

Works this paper leans on

74 extracted references · 74 canonical work pages · cited by 1 Pith paper · 5 internal anchors

  1. [1]

    Lazarov, and Ole Sigmund

    Niels Aage, Erik Andreassen, Boyan S. Lazarov, and Ole Sigmund. Giga-voxel computational morphogenesis for structural design.Nature, 550(7674):84–86, 2017. doi: 10.1038/nature23911

    work page doi:10.1038/nature23911 2017

  2. [2]

    Abueidda, Seid Koric, and Nahed A

    Diab W. Abueidda, Seid Koric, and Nahed A. Sobh. Topology optimization of 2D structures with nonlinearities using deep learning.Computers & Structures, 237:106283, 2020. doi: 10.1016/j.compstruc.2020.106283

    work page doi:10.1016/j.compstruc.2020.106283 2020

  3. [3]

    Automated dynamic algorithm configuration.Journal of Artificial Intelligence Research, 75:1633–1699, 2022

    Steven Adriaensen, Andr´ e Biedenkapp, Gresa Shala, Noor Awad, Theresa Eimer, Marius Lin- dauer, and Frank Hutter. Automated dynamic algorithm configuration.Journal of Artificial Intelligence Research, 75:1633–1699, 2022. doi: 10.1613/jair.1.13922

    work page doi:10.1613/jair.1.13922 2022

  4. [4]

    Lazarov, and Ole Sig- mund

    Erik Andreassen, Anders Clausen, Mattias Schevenels, Boyan S. Lazarov, and Ole Sigmund. Efficient topology optimization in MATLAB using 88 lines of code.Structural and Multidisci- plinary Optimization, 43(1):1–16, 2011. doi: 10.1007/s00158-010-0594-7

    work page doi:10.1007/s00158-010-0594-7 2011

  5. [5]

    The Claude model card and evaluations

    Anthropic. The Claude model card and evaluations. Anthropic System Cards, 2024. https: //www.anthropic.com/system-cards

    work page 2024

  6. [6]

    Olson, Jacob Schroder, and Ben Southworth

    Nathan Bell, Luke N. Olson, Jacob Schroder, and Ben Southworth. PyAMG: Algebraic multigrid solvers in Python.Journal of Open Source Software, 8(87):5495, 2023. doi: 10.21105/joss.05495

    work page doi:10.21105/joss.05495 2023

  7. [7]

    Martin P. Bendsøe. Optimal shape design as a material distribution problem.Structural Optimization, 1(4):193–202, 1989. doi: 10.1007/BF01650949

    work page doi:10.1007/bf01650949 1989

  8. [8]

    Bendsøe and Noboru Kikuchi

    Martin P. Bendsøe and Noboru Kikuchi. Generating optimal topologies in structural design using a homogenization based method.Computer Methods in Applied Mechanics and Engi- neering, 71(2):197–224, 1988. doi: 10.1016/0045-7825(88)90086-2

    work page doi:10.1016/0045-7825(88)90086-2 1988

  9. [9]

    Bendsøe and Ole Sigmund

    Martin P. Bendsøe and Ole Sigmund. Material interpolation schemes in topology optimization. Archive of Applied Mechanics, 69(9–10):635–654, 1999. doi: 10.1007/s004190050248. 30

    work page doi:10.1007/s004190050248 1999

  10. [10]

    Curriculum learning

    Yoshua Bengio, J´ erˆ ome Louradour, Ronan Collobert, and Jason Weston. Curriculum learning. InProceedings of the 26th Annual International Conference on Machine Learning (ICML), pages 41–48, 2009. doi: 10.1145/1553374.1553380

    work page doi:10.1145/1553374.1553380 2009

  11. [11]

    Machine learning for combinatorial optimization: A methodological tour d’horizon.European Journal of Operational Research, 290(2):405–421, 2021

    Yoshua Bengio, Andrea Lodi, and Antoine Prouvost. Machine learning for combinatorial optimization: A methodological tour d’horizon.European Journal of Operational Research, 290(2):405–421, 2021. doi: 10.1016/j.ejor.2020.07.063

    work page doi:10.1016/j.ejor.2020.07.063 2021

  12. [12]

    Algorithms for hyper- parameter optimization

    James Bergstra, R´ emi Bardenet, Yoshua Bengio, and Bal´ azs K´ egl. Algorithms for hyper- parameter optimization. InAdvances in Neural Information Processing Systems, volume 24, pages 2546–2554, 2011

    work page 2011

  13. [13]

    Furkan Bozkurt, Theresa Eimer, Frank Hutter, and Marius Lindauer

    Andr´ e Biedenkapp, H. Furkan Bozkurt, Theresa Eimer, Frank Hutter, and Marius Lindauer. Dynamic algorithm configuration: Foundation of a new meta-algorithmic framework. InPro- ceedings of the 24th European Conference on Artificial Intelligence (ECAI 2020), pages 427– 434, 2020. doi: 10.3233/FAIA200122

    work page doi:10.3233/faia200122 2020

  14. [14]

    Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges.WIREs Data Mining and Knowledge Discovery, 13(2):e1484, 2023

    Bernd Bischl, Martin Binder, Michel Lang, Tobias Pielok, Jakob Richter, Stefan Coors, Janek Thomas, Theresa Ullmann, Marc Becker, Anne-Laure Boulesteix, Difan Deng, and Marius Lindauer. Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges.WIREs Data Mining and Knowledge Discovery, 13(2):e1484, 2023. doi: 10.1002/ widm.1484

    work page 2023

  15. [15]

    C., Barends, R., Biswas, R., Boixo, S., Brandao, F

    Daniil A. Boiko, Robert MacKnight, Ben Kline, and Gabe Gomes. Autonomous chemical research with large language models.Nature, 624(7992):570–578, 2023. doi: 10.1038/s41586- 023-06792-0

    work page doi:10.1038/s41586- 2023

  16. [16]

    Filters in topology optimization.International Journal for Numerical Methods in Engineering, 50(9):2143–2158, 2001

    Blaise Bourdin. Filters in topology optimization.International Journal for Numerical Methods in Engineering, 50(9):2143–2158, 2001. doi: 10.1002/nme.116

    work page doi:10.1002/nme.116 2001

  17. [17]

    Brown, Anthony P

    Nathan K. Brown, Anthony P. Garland, Georges M. Fadel, and Gang Li. Deep reinforcement learning for engineering design through topology optimization of elementally discretized design domains.Materials & Design, 218:110672, 2022. doi: 10.1016/j.matdes.2022.110672

    work page doi:10.1016/j.matdes.2022.110672 2022

  18. [18]

    Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhari- wal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agar- wal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Li...

    work page 1901

  19. [19]

    T. E. Bruns and D. A. Tortorelli. Topology optimization of non-linear elastic structures and compliant mechanisms.Computer Methods in Applied Mechanics and Engineering, 190(26–27): 3443–3459, 2001. doi: 10.1016/S0045-7825(00)00278-4

    work page doi:10.1016/s0045-7825(00)00278-4 2001

  20. [20]

    One-shot generation of near-optimal topology through theory-driven machine learning.Computer-Aided Design, 109:12–21, 2019

    Ruijin Cang, Hope Yao, and Yi Ren. One-shot generation of near-optimal topology through theory-driven machine learning.Computer-Aided Design, 109:12–21, 2019. doi: 10.1016/ j.cad.2018.12.008

    work page 2019

  21. [21]

    TOuNN: Topology optimization using neural networks.Structural and Multidisciplinary Optimization, 63(3):1135–1149, 2021

    Aaditya Chandrasekhar and Krishnan Suresh. TOuNN: Topology optimization using neural networks.Structural and Multidisciplinary Optimization, 63(3):1135–1149, 2021. doi: 10.1007/ s00158-020-02748-4. 31

    work page 2021

  22. [22]

    Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Pond´ e de Oliveira Pinto, Jared Kaplan, Harrison Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri, Gretchen Krueger, Michael Petrov, Heidy Khlaaf, Girish Sastry, Pamela Mishkin, Brooke Chan, Scott Gray, Nick Ryder, Mikhail Pavlov, Alethea Power, Lukasz Kaiser, Mo- hammad B...

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  23. [23]

    Deaton and Ramana V

    Joshua D. Deaton and Ramana V. Grandhi. A survey of structural and multidisciplinary continuum topology optimization: post 2000.Structural and Multidisciplinary Optimization, 49(1):1–38, 2014. doi: 10.1007/s00158-013-0956-z

    work page doi:10.1007/s00158-013-0956-z 2000

  24. [24]

    Self-directed online machine learning for topology optimization.Nature Communications, 13(1):388, 2022

    Changyu Deng, Yizhou Wang, Can Qin, Yun Fu, and Wei Lu. Self-directed online machine learning for topology optimization.Nature Communications, 13(1):388, 2022. doi: 10.1038/ s41467-021-27713-7

    work page 2022

  25. [25]

    Danny Driess, Fei Xia, Mehdi S. M. Sajjadi, Corey Lynch, Aakanksha Chowdhery, Brian Ichter, Ayzaan Wahid, Jonathan Tompson, Quan Vuong, Tianhe Yu, Wenlong Huang, Yevgen Cheb- otar, Pierre Sermanet, Daniel Duckworth, Sergey Levine, Vincent Vanhoucke, Karol Hausman, Marc Toussaint, Klaus Greff, Andy Zeng, Igor Mordatch, and Pete Florence. PaLM-E: An em- bod...

    work page 2023

  26. [26]

    Automatic projection parameter increase for three-field density-based topology optimization.Structural and Multidisciplinary Optimization, 68(2):33,

    Peter Dunning and Fabian Wein. Automatic projection parameter increase for three-field density-based topology optimization.Structural and Multidisciplinary Optimization, 68(2):33,

    work page

  27. [27]

    doi: 10.1007/s00158-025-03968-2

    work page doi:10.1007/s00158-025-03968-2

  28. [28]

    DACBench: A benchmark library for dynamic algorithm configuration

    Theresa Eimer, Andr´ e Biedenkapp, Maximilian Reimer, Steven Adriaensen, Frank Hutter, and Marius Lindauer. DACBench: A benchmark library for dynamic algorithm configuration. In Proceedings of the 30th International Joint Conference on Artificial Intelligence (IJCAI-21), pages 1668–1674, 2021. doi: 10.24963/ijcai.2021/230

    work page doi:10.24963/ijcai.2021/230 2021

  29. [29]

    A new generation 99 line Matlab code for compliance topology optimization and its extension to 3D.Structural and Multidisciplinary Optimization, 62:2211–2228, 2020

    Federico Ferrari and Ole Sigmund. A new generation 99 line Matlab code for compliance topology optimization and its extension to 3D.Structural and Multidisciplinary Optimization, 62:2211–2228, 2020. doi: 10.1007/s00158-020-02629-w

    work page doi:10.1007/s00158-020-02629-w 2020

  30. [30]

    K., Prévost, J

    James K. Guest, Jean H. Pr´ evost, and Ted Belytschko. Achieving minimum length scale in topology optimization using nodal design variables and projection functions.International Journal for Numerical Methods in Engineering, 61(2):238–254, 2004. doi: 10.1002/nme.1064

    work page doi:10.1002/nme.1064 2004

  31. [31]

    Guest, Alireza Asadpoure, and Seung-Hyun Ha

    James K. Guest, Alireza Asadpoure, and Seung-Hyun Ha. Eliminating beta-continuation from Heaviside projection and density filter algorithms.Structural and Multidisciplinary Optimiza- tion, 44(4):443–453, 2011. doi: 10.1007/s00158-011-0676-1

    work page doi:10.1007/s00158-011-0676-1 2011

  32. [32]

    Introducing Gemini 2.0: Our new AI model for the agentic era

    Demis Hassabis and Koray Kavukcuoglu. Introducing Gemini 2.0: Our new AI model for the agentic era. Google DeepMind Blog, December 2024. https://blog.google/technology/google- deepmind/google-gemini-ai-update-december-2024/. 32

    work page 2024

  33. [33]

    Reinforcement learning and graph embedding for binary truss topology optimization under stress and displacement constraints.Frontiers in Built Environment, 6:59, 2020

    Kazuki Hayashi and Makoto Ohsaki. Reinforcement learning and graph embedding for binary truss topology optimization under stress and displacement constraints.Frontiers in Built Environment, 6:59, 2020. doi: 10.3389/fbuil.2020.00059

    work page doi:10.3389/fbuil.2020.00059 2020

  34. [34]

    Neural reparameterization im- proves structural optimization

    Stephan Hoyer, Jascha Sohl-Dickstein, and Sam Greydanus. Neural reparameterization im- proves structural optimization. Technical Report arXiv:1909.04240, arXiv, 2019. arXiv preprint

    work page arXiv 1909

  35. [35]

    Hoos, and Kevin Leyton-Brown

    Frank Hutter, Holger H. Hoos, and Kevin Leyton-Brown. Sequential model-based op- timization for general algorithm configuration. InLearning and Intelligent Optimization (LION 5), volume 6683 ofLecture Notes in Computer Science, pages 507–523, 2011. doi: 10.1007/978-3-642-25566-3 40

    work page doi:10.1007/978-3-642-25566-3 2011

  36. [36]

    Springer, Cham, 2019

    Frank Hutter, Lars Kotthoff, and Joaquin Vanschoren, editors.Automated Machine Learning: Methods, Systems, Challenges. Springer, Cham, 2019. doi: 10.1007/978-3-030-05318-5

    work page doi:10.1007/978-3-030-05318-5 2019

  37. [37]

    Population Based Training of Neural Networks

    Max Jaderberg, Valentin Dalibard, Simon Osindero, Wojciech M. Czarnecki, Jeff Donahue, Ali Razavi, Oriol Vinyals, Tim Green, Iain Dunning, Karen Simonyan, Chrisantha Fernando, and Koray Kavukcuoglu. Population based training of neural networks. Technical Report arXiv:1711.09846, arXiv, 2017. arXiv preprint

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  38. [38]

    Kallioras, Georgios Kazakis, and Nikos D

    Nikolaos Ath. Kallioras, Georgios Kazakis, and Nikos D. Lagaros. Accelerated topology op- timization by means of deep learning.Structural and Multidisciplinary Optimization, 62(3): 1185–1212, 2020. doi: 10.1007/s00158-020-02545-z

    work page doi:10.1007/s00158-020-02545-z 2020

  39. [39]

    Lazarov and Ole Sigmund

    Boyan S. Lazarov and Ole Sigmund. Filters in topology optimization based on Helmholtz-type differential equations.International Journal for Numerical Methods in Engineering, 86(6): 765–781, 2011. doi: 10.1002/nme.3072

    work page doi:10.1002/nme.3072 2011

  40. [40]

    Lazarov, Fengwen Wang, and Ole Sigmund

    Boyan S. Lazarov, Fengwen Wang, and Ole Sigmund. Length scale and manufacturability in density-based topology optimization.Archive of Applied Mechanics, 86(1):189–218, 2016. doi: 10.1007/s00419-015-1106-4

    work page doi:10.1007/s00419-015-1106-4 2016

  41. [41]

    Hy- perband: A novel bandit-based approach to hyperparameter optimization.Journal of Machine Learning Research, 18(185):1–52, 2018

    Lisha Li, Kevin Jamieson, Giulia DeSalvo, Afshin Rostamizadeh, and Ameet Talwalkar. Hy- perband: A novel bandit-based approach to hyperparameter optimization.Journal of Machine Learning Research, 18(185):1–52, 2018

    work page 2018

  42. [42]

    Critical current of a Josephson junction containing a conical magnet

    Shengcai Liu, Caishun Chen, Xinghua Qu, Ke Tang, and Yew-Soon Ong. Large language models as evolutionary optimizers. In2024 IEEE Congress on Evolutionary Computation (CEC), pages 1–8, 2024. doi: 10.1109/CEC60901.2024.10611913

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1109/cec60901.2024.10611913 2024

  43. [43]

    Self-Refine: Iterative refinement with self-feedback

    Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegr- effe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, Shailendra Gupta, Bod- hisattwa Prasad Majumder, Katherine Hermann, Sean Welleck, Amir Yazdanbakhsh, and Peter Clark. Self-Refine: Iterative refinement with self-feedback. InAdvances in Neural Infor- mation Proces...

    work page 2023

  44. [44]

    Diffusion models beat GANs on topology optimization

    Fran¸ cois Maz´ e and Faez Ahmed. Diffusion models beat GANs on topology optimization. Proceedings of the AAAI Conference on Artificial Intelligence, 37(8):9108–9116, 2023. doi: 10.1609/aaai.v37i8.26093. 33

    work page doi:10.1609/aaai.v37i8.26093 2023

  45. [45]

    Zhenguo Nie, Tong Lin, Haoliang Jiang, and Levent B. Kara. TopologyGAN: Topology opti- mization using generative adversarial networks based on physical fields over the initial domain. ASME Journal of Mechanical Design, 143(3):031715, 2021. doi: 10.1115/1.4049533

    work page doi:10.1115/1.4049533 2021

  46. [46]

    GPT-4 Technical Report

    OpenAI. GPT-4 technical report. Technical Report arXiv:2303.08774, OpenAI, 2023. arXiv preprint

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  47. [47]

    Bernstein

    Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. Generative agents: Interactive simulacra of human behavior. In ACM Symposium on User Interface Software and Technology, pages 1–22, 2023. doi: 10.1145/ 3586183.3606763

    work page arXiv 2023

  48. [48]

    Critical current of a Josephson junction containing a conical magnet

    Thiago Rios, Stefan Menzel, Michael Wong, and Yew-Soon Ong. Large language and text- to-3D models for engineering design optimization. In2024 IEEE Congress on Evolutionary Computation (CEC), pages 1–8, 2024. doi: 10.1109/CEC60901.2024.10612061

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1109/cec60901.2024.10612061 2024

  49. [49]

    Automatic penalty continuation in structural topology optimization.Structural and Multidisciplinary Optimization, 52(6):1205–1221, 2015

    Susana Rojas-Labanda and Mathias Stolpe. Automatic penalty continuation in structural topology optimization.Structural and Multidisciplinary Optimization, 52(6):1205–1221, 2015. doi: 10.1007/s00158-015-1277-1

    work page doi:10.1007/s00158-015-1277-1 2015

  50. [50]

    Nature (2023)

    Bernardino Romera-Paredes, Mohammadamin Barekatain, Alexander Novikov, Matej Balog, M Pawan Kumar, Emilien Dupont, Francisco JR Ruiz, Jordan S Ellenberg, Pengming Wang, Omar Fawzi, et al. Mathematical discoveries from program search with large language models. Nature, 625(7995):468–475, 2024. doi: 10.1038/s41586-023-06924-6

    work page doi:10.1038/s41586-023-06924-6 2024

  51. [51]

    Toolformer: Language models can teach themselves to use tools

    Timo Schick, Jane Dwivedi-Yu, Roberto Dess` ı, Roberta Raileanu, Maria Lomeli, Luke Zettle- moyer, Nicola Cancedda, and Thomas Scialom. Toolformer: Language models can teach themselves to use tools. InAdvances in Neural Information Processing Systems, volume 36, pages 2510–2544, 2023

    work page 2023

  52. [52]

    Adams, and Nando de Freitas

    Bobak Shahriari, Kevin Swersky, Ziyu Wang, Ryan P. Adams, and Nando de Freitas. Taking the human out of the loop: A review of Bayesian optimization.Proceedings of the IEEE, 104 (1):148–175, 2016. doi: 10.1109/JPROC.2015.2494218

    work page doi:10.1109/jproc.2015.2494218 2016

  53. [53]

    Learning step-size adaptation in CMA-ES

    Gresa Shala, Andr´ e Biedenkapp, Noor Awad, Steven Adriaensen, Marius Lindauer, and Frank Hutter. Learning step-size adaptation in CMA-ES. InParallel Problem Solving from Nature (PPSN XVI), volume 12269 ofLecture Notes in Computer Science, pages 691–706, 2020

    work page 2020

  54. [54]

    Topology optimization via machine learning and deep learning: a review.Journal of Computational Design and Engineering, 10(4):1736–1766, 2023

    Seungyeon Shin, Dongju Shin, and Namwoo Kang. Topology optimization via machine learning and deep learning: a review.Journal of Computational Design and Engineering, 10(4):1736– 1766, 2023. doi: 10.1093/jcde/qwad072

    work page doi:10.1093/jcde/qwad072 2023

  55. [55]

    Reflexion: Language agents with verbal reinforcement learning

    Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. Reflexion: Language agents with verbal reinforcement learning. InAdvances in Neural Infor- mation Processing Systems, volume 36, pages 8634–8652, 2023

    work page 2023

  56. [56]

    A 99 line topology optimization code written in Matlab.Structural and Multidisciplinary Optimization, 21(2):120–127, 2001

    Ole Sigmund. A 99 line topology optimization code written in Matlab.Structural and Multi- disciplinary Optimization, 21(2):120–127, 2001. doi: 10.1007/s001580050176

    work page doi:10.1007/s001580050176 2001

  57. [57]

    Morphology-based black and white filters for topology optimization.Structural and Multidisciplinary Optimization, 33(4-5):401–424, 2007

    Ole Sigmund. Morphology-based black and white filters for topology optimization.Structural and Multidisciplinary Optimization, 33(4-5):401–424, 2007. doi: 10.1007/s00158-006-0087-x. 34

    work page doi:10.1007/s00158-006-0087-x 2007

  58. [58]

    Topology optimization approaches.Structural and Multi- disciplinary Optimization, 48(6):1031–1055, 2013

    Ole Sigmund and Kurt Maute. Topology optimization approaches.Structural and Multidisci- plinary Optimization, 48(6):1031–1055, 2013. doi: 10.1007/s00158-013-0978-6

    work page doi:10.1007/s00158-013-0978-6 2013

  59. [59]

    Numerical instabilities in topology optimization: a survey on procedures dealing with checkerboards, mesh-dependencies and local minima

    Ole Sigmund and Joakim Petersson. Numerical instabilities in topology optimization: a survey on procedures dealing with checkerboards, mesh-dependencies and local minima.Structural Optimization, 16(1):68–75, 1998. doi: 10.1007/BF01214002

    work page doi:10.1007/bf01214002 1998

  60. [60]

    Jasper Snoek, Hugo Larochelle, and Ryan P. Adams. Practical Bayesian optimization of ma- chine learning algorithms. InAdvances in Neural Information Processing Systems, volume 25, pages 2951–2959, 2012

    work page 2012

  61. [61]

    Neural networks for topology optimization.Russian Journal of Numerical Analysis and Mathematical Modelling, 34(4):215–223, 2019

    Ivan Sosnovik and Ivan Oseledets. Neural networks for topology optimization.Russian Journal of Numerical Analysis and Mathematical Modelling, 34(4):215–223, 2019. doi: 10.1515/rnam- 2019-0018

    work page doi:10.1515/rnam- 2019

  62. [62]

    On the trajectories of penalization methods for topol- ogy optimization.Structural and Multidisciplinary Optimization, 21(2):128–139, 2001

    Mathias Stolpe and Krister Svanberg. On the trajectories of penalization methods for topol- ogy optimization.Structural and Multidisciplinary Optimization, 21(2):128–139, 2001. doi: 10.1007/s001580050177

    work page doi:10.1007/s001580050177 2001

  63. [63]

    Svanberg, ‘The method of moving asymptotes—a new method for structural optimization’, Numerical Meth En- gineering, vol

    Krister Svanberg. The method of moving asymptotes — a new method for structural opti- mization.International Journal for Numerical Methods in Engineering, 24(2):359–373, 1987. doi: 10.1002/nme.1620240207

    work page doi:10.1002/nme.1620240207 1987

  64. [64]

    A class of globally convergent optimization methods based on conservative convex separable approximations.SIAM Journal on Optimization, 12(2):555–573, 2002

    Krister Svanberg. A class of globally convergent optimization methods based on conservative convex separable approximations.SIAM Journal on Optimization, 12(2):555–573, 2002. doi: 10.1137/S1052623499362822

    work page doi:10.1137/s1052623499362822 2002

  65. [65]

    Lazarov, and Ole Sigmund

    Fengwen Wang, Boyan S. Lazarov, and Ole Sigmund. On projection methods, convergence and robust formulations in topology optimization.Structural and Multidisciplinary Optimization, 43(6):767–784, 2011. doi: 10.1007/s00158-010-0602-y

    work page doi:10.1007/s00158-010-0602-y 2011

  66. [66]

    Voyager: An open-ended embodied agent with large lan- guage models.Transactions on Machine Learning Research, pages 1–27, 2024

    Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. Voyager: An open-ended embodied agent with large lan- guage models.Transactions on Machine Learning Research, pages 1–27, 2024. https: //openreview.net/forum?id=ehfRiF0R3a

    work page 2024

  67. [67]

    Chain-of-thought prompting elicits reasoning in large language models

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed Chi, Quoc Le, and Denny Zhou. Chain-of-thought prompting elicits reasoning in large language models. InAdvances in Neural Information Processing Systems, volume 35, pages 24824–24837, 2022

    work page 2022

  68. [68]

    Woldseth, Niels Aage, Jakob Andreas Bærentzen, and Ole Sigmund

    Rebekka V. Woldseth, Niels Aage, Jakob Andreas Bærentzen, and Ole Sigmund. On the use of artificial neural networks in topology optimisation.Structural and Multidisciplinary Optimization, 65(10):294, 2022. doi: 10.1007/s00158-022-03347-1

    work page doi:10.1007/s00158-022-03347-1 2022

  69. [69]

    Evolutionary com- putation in the era of large language model: Survey and roadmap.IEEE Transactions on Evolutionary Computation, 29(2):534–554, 2024

    Xingyu Wu, Sheng-hao Wu, Jibin Wu, Liang Feng, and Kay Chen Tan. Evolutionary com- putation in the era of large language model: Survey and roadmap.IEEE Transactions on Evolutionary Computation, 29(2):534–554, 2024. doi: 10.1109/TEVC.2024.3506731

    work page doi:10.1109/tevc.2024.3506731 2024

  70. [70]

    Le, Denny Zhou, and Xinyun Chen

    Chengrun Yang, Xuezhi Wang, Yifeng Lu, Hanxiao Liu, Quoc V. Le, Denny Zhou, and Xinyun Chen. Large language models as optimizers. InThe Twelfth International Conference on Learning Representations, 2024. https://openreview.net/forum?id=Bb4VGOWELI. 35

    work page 2024

  71. [71]

    ReAct: Synergizing reasoning and acting in language models

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. ReAct: Synergizing reasoning and acting in language models. InInternational Conference on Learning Representations, 2023. https://openreview.net/forum?id=WE vluYUL-X

    work page 2023

  72. [72]

    ReEvo: Large language models as hyper-heuristics with reflective evolution

    Haoran Ye, Jiarui Wang, Zhiguang Cao, Federico Berto, Chuanbo Hua, Haeyeon Kim, Jinkyoo Park, and Guojie Song. ReEvo: Large language models as hyper-heuristics with reflective evolution. InAdvances in Neural Information Processing Systems, 2024. https://proceedings.neurips.cc/paper files/paper/2024/hash/ 4ced59d480e07d290b6f29fc8798f195-Abstract-Conference.html

    work page 2024

  73. [73]

    Deep learning for determining a near-optimal topological design without any iteration.Structural and Multidisciplinary Optimization, 59(3):787–799, 2019

    Yonggyun Yu, Taeil Hur, Jaeho Jung, and In Gwun Jang. Deep learning for determining a near-optimal topological design without any iteration.Structural and Multidisciplinary Opti- mization, 59(3):787–799, 2019. doi: 10.1007/s00158-018-2101-5

    work page doi:10.1007/s00158-018-2101-5 2019

  74. [74]

    Zhou and G

    M. Zhou and G. I. N. Rozvany. The COC algorithm, Part II: Topological, geometrical and generalized shape optimization.Computer Methods in Applied Mechanics and Engineering, 89 (1–3):309–336, 1991. doi: 10.1016/0045-7825(91)90046-9. 36

    work page doi:10.1016/0045-7825(91)90046-9 1991